Paul J Van Diest, Professor and Head, Department of Pathology, University Medical Centre Utrecht advocated for the wider implementation of AI in digital pathology. He estimated that only 0.28% of the total costs of healthcare in the Netherlands is being spent on pathology. Even with such little spending, scientist can still get so much out of it.
Van Diest has implemented 14 algorithms and is using them routinely some on specific demand in clinical practice. The cost of implementing these algorithms was around 500,000 euros. AI stains also help save costs as AI requires fewer stains in comparison to immune stains.
Unfortunately, commercial AI algorithm implementation is expensive, and most labs cannot afford it. Van Diest added that insurance companies rarely reimburse AI-based diagnostics, creating a financial disconnect.
A trial comparing AI diagnosis and non-AI supported diagnosis in breast, prostate and colon cancer to see the cost that could be saved by doing fewer stains. The prostate trial saved 2400 euros.
Joint regional or national AI systems, such as shared servers, can reduce implementation costs and make AI more accessible to multiple labs. Van Diest concluded that AI is increasingly necessary to address shortages of pathologists and technicians, helping maintain efficiency and quality in clinical practice.
